Discontinuous Al-SiC Composites Formed by a Lovv Cost Chemically Activated Infiltration Technique Pridobivanje in kemijska infiltracija poroznih SiC vzorcev z Al-Si talino V. M. Kevorkijan1, zasebni raziskovalec, Maribor, Slovenija Prejem rokopisa - received: 1996-10-01; sprejem za objavo - accepted for publication: 1996-11-04 In this vvork, the preparation of porous SiC preforms from SiC particles, piatelets and whiskers have been demonstarted. Near net shape preforms, prepared by vacuum casting, vvere sintered and then covered by S1O2 layer using a cost effective oxidation in air at 1175 K for 10h. Surface engineered SiC preforms vvere than pressureless infiltrated in nitrogen atmosphere (96 vol% N2 + 4 vol% Ar) by a Al-Si melt containing 0.5-3 wt% Mg. Based on this, a mathematical model of spontaneous infiltration of a porous ceramic preform has been suggested. The roie of magnesium and nitrogen atmosphere vvas quantitatively evaluated among the other important processing parameters (porosity of preform, the specific surface area, etc.) collected in a new term named preform infiltrability. Moreover, the influence of the above listed parameters on the infiltration rate (expressed as infiltration length and function of time) has also been demonstrated. The optimal conditions for spontaneous and cost effective pressureless infiltration of porous SiC preforms by molten aluminium alloy has been selected and experimentaty confirmed. Key vvords: porous SiC preforms, vacuum casting, pressureless infiltration, infiltration kinetics, infiltrability of porous preforms V delu je opisana izdelava poroznih SiC vzorcev z vakuumskim vlivanjem in njihovo sintranje do poroznih predoblik, sestavljenih iz SiC delcev različne oblike: okrogli, heksagonalne ploščice in kratka vlakna. Z oksidacijo na zraku smo prevlekli površino poroznih SiC predoblik s tanko plastjo S1O2 in izboljšali omočljivost med SiC in Al talino. V nadaljnjem delu smo infiltriraii porozne keramične vzorce z Al-Si-Mg talino v dušikovi atmosferi (96 vol% N2 + 4 vol% Ar) pri normalnem tlaku. Na podlagi pridobljenih rezultatov smo razvili matematični model infiltracije, ki opisuje kinetiko procesa v funkciji poroznosti in specifične površine pripravljenih poroznih vzorcev, vsebnosti dušika v atmosferi in sestave Al zlitine. Model je osnova za nadaljnji razvoj tehnologije priprave Al-SiC kompozitov s spontano oz. nizkotlačno infiltracijo poroznih SiC vzorcev. Ključne besede: porozne SiC predoblike, vakuumsko vlivanje, spontana infiltracija, kinetika infiltracije, infiltrabilnost poroznih predoblik 1 Introduction The need for high strength. lightweight, and high stiffness materials has, in recent years, attracted much in-terest to the development of the manufacturing processes of metal matrix composites (MMCs)1. The most important limitation of the fabrication of MMCs by liquid-phase processes resides in the compatibility betvveen the reinforcement and the matrix2. This compatibility is par-ticularly important in the čase of aluminium-based composites, because Al is usually covered vvith a thin oxide layer that prevents vvetting, and vvhen uncovered, it read-ily reacts vvith most ceramics to form intermetalics. In particular, liquid aluminium reacts vvith SiC to produce aluminium carbide and free silicon. Wettability and reac-tivity determine the quality of the bond betvveen both materials and, therefore, greatly influence the final properties of the composite. In many instances the properties of a reinforced metal have been shovvn to provide a performance advantage over a monolithic metal, but the high cost of producing the composite has prohibited vvidespread commercial use. Liquid-metal processes have the potential to be more economical; hovever, the non-vvetting nature of many ceramics by molten aluminium, vvhich results in 1 Dr. Varužan M. KEVORKIJAN Lackova 139 2341 Limbuš. Slovenija poor ceramic/metal interfaces and incomplete infiltration, has been an obstacle. Melt infiltration is a popular technique for fabricating MMCs, as it allovv near-net shape fabrication of components and material vvith a high reinforcing phase content. The molten metal may penetrate the porous preforms either under the action of an external force (pressure casting3 and vacuum assisted Iiquid infiltration process4) or through a capillary pressure vvhich is created once the molten metal vvets the ceramic surface (pressureless infiltration5). Several pressure casting methods have been used for preparing MMCs. The operating principle of a hydro-static pressure infiltration device6 is to use pressurised gas to force molten metal into an evacuated die. Another pressure casting technique is relatively simple7: pre-heat-ing the particle aggregate in a special mould and then adding 3 MPa pressure to the molten metal poured on the particle aggregate so as to encourage penetration vvhich results in a metal-particle composite. Recently, a bottom mixing process has also been suggested, vvhere an evacuated packed bed in the bottom of a crucible is covered vvith a melt, and than stirrer shears the interface betvveen the particles and the melt, resulting in incorporation8. Different fabrication methods using vacuum tech-niques for cast-in-place hardfacing of casting vvere also described9. In these processes, aluminium poured into a sand mould is dravvn by vacuum into a porous layer of reinforcing phase (named - preform) placed on a wall of the mould cavity. A recent molten metal process is the Lanxide Corp. Primex™ pressureless infiltration process10 ". In this process a packed bed of ceramic powder is infiltrated by an Al-Mg alloy, without any applied pressure, in a nitrogen atmosphere. The resulting composite, vvhich has a packed bed density of about 55 vol.-%, can than be di-luted in the appropriate matrix alloy. Ceramic particles of SiC and AI2O3, vvith particle size as fine as about 1 pm have been infiltrated in this way, and at infiltration rates of up to the order of centimetres per minute under specific processing conditions. Processing details of the Primex™ route are proprietary, but it vvould appear to be a very competitive process for higher volume fraction composites. The Lanxide Corporation has made exten-sive efforts to protect this very valuable technology and has vvell over 100 U.S. patents and over 1500 foreign patents pending, vvith nearly 50 U.S. patents and over 100 foreign patents being issued or allovved by the mid-dle of 1989. In this vvork, the preparation of porous SiC preforms made by SiC particles, platelets and vvhiskers have been demonstrated. The surface of SiC preforms has been cov-ered by SiCb layer using cost effective oxidation in air. Chemically treated preforms vvere than pressurelless infiltrated by an Al-Mg alloy in nitrogen atmosphere. The conditions for spontaneous (as used, spontaneously means vvithout the aid of any externally applied pressure or vacuum), pressureless infiltration, vvhich include the use of a magnesium containing alloy and a nitrogenous atmosphere have been already vvell documented in literature, by the inventors12. Hovvever, the offered explana-tion is semi-empirical based on the vvell knovvn role of magnesium vvhich decreases the surface tension of a molten aluminium alloy. As stated by inventors12, this alone does not induce spontaneous infiltration, but a nitrogen atmosphere may cause a further reduction in the surface tension, thus promoting vvetting. Additionally, the reactivity of magnesium induces interfacial reactions vvith solid ceramic surfaces. These reactions typically are not sufficient to promote spontaneous vvetting, but again in combination vvith a nitrogen atmosphere they may change or be altered, thus allovving the observed infiltration. These results clearly demonstrated that the combination of magnesium in the alloy and a nitrogeneous atmosphere leads to the spontaneous infiltration of aluminium alloy into ceramic fillers. Hovvever, little in-formation is available on the effect of a nitrogen atmosphere on vvetting. Some authors13 found that vvhen fabri-cating aluminium alloy matrix composites via compocasting, the use of a nitrogen atmosphere and a bubble-degassing step vvith nitrogen yielded composites vvith much lovver porosity than those produced similarly vvith argon, but these results may not be assciated vvith enhanced vvetting. In the present paper a mathematical model of spontaneous infiltration of a porous ceramic preform has been suggested. The role of magnesium and nitrogenous atmosphere vvas quantitatively evaluated among the impor-tant processing parameters (porosity of the preform, the specific surface area, surface tension and the contact angle). Moreover, the influence of above listed parameters on the infiltration rate (expressed by the infiltration length as a function of time) has been also demonstrated. In this way, the optimal conditions for spontaneous and cost effective pressureless infiltration of porous SiC preforms by molten aluminium alloy vvere selected and ex-perimentaly confirmed. 2 Pressureless infiltration - theoretical considera-tions A. Capilllar}> Law Spontaneous infiltration of a liquid into a porous medium takes plače vvhen the liquid vvets the solid. Other-vvise, a minimum external pressure should be applied. This threshold pressure P (also called capillary pressure) is related to the contact angle 0 and the particle size through the so-called capillary law or Laplace equation: P = 6^ylv cos0 Vp/((l-Vp)D) (1) vvhere y\v is the liquid-vapor surface tension, X a factor vvhich depends on the geometry of the particles, D the mean diameter of the particles, and Vp the particulate volume fraction. Note that product (- yiv cos 0) is the vvork of immersion Wi defined as the change in the free energy on immersing the solid in the liquid. The vvork of immersion can be vvritten in terms of the threshold (or capillary) pressure through the follovving expression: W, = P(l-Vp)/SsPVp (2) vvhere Ss is the specific surface area (the surface area per unit mass of porous preform) and p is the density of the particulate. Unfortunately, the Laplace equation de-scribes the situation for a cylindrical tube, a very crude model for the types of porous media under considera-tion here. This model, for example, cannot be applied to irregularly shaped pores vvhere the effect of both pore geometry and netvvork cooperatively combine vvith contact angle hysteresis14. Hovvever, White15 derived a spe-cialized expression based on the Laplace equation relat-ing the pressure, P required to prevent capillary rise in porous media for vvhich the specific surface area Ss, solid density p, surface tension yiv, contact angle 0, and porosity a, are knovvn: P = (l-£) p Ssylv cos0 /e (3) B. Darcy's Law The flovv of an incompressible fluid through a porous medium is governed by Darcy's lavv16. For unidirectional flavv, and neglecting any effect of gravity, Darcy's lavv can be vvritten as v0 = - (k/fi) (dP/dx) (4) where v0 is the superficial velocity of the fluid (the ve-loeity of the fluid as measured by the volumetric flovv rate per unit cross sectional area vvhere the cross section is taken perpendicular to the average direction of flow), ji the viscosity of the liquid, dP/dx the pressure gradient at the infiltration front, and k the intrinsic permeability. It has been found empirically that the intrinsic perme-ability k of a porous medium is proportional to the square of the mean particulate diameter17 k = aD2 (5) where the constant a must be determined experimen-tally. The superficial velocity v0 can be related to the actual velocity in the porous medium (dx/dt) by means of the particulate volume fraction Vp: v0 = (1-Vp) dx/dt (6) Combining Eqs. (4) and (6) and integrating, the ex-pression for the infiltration length, L as a function of time and the pressure drop in the liquid metal #9P can be written as: L = [2ktAP/)d(l-V )]' On the other hand, for pressureless infiltration the pressure drop should be at least equal to the threshold (or capillary) pressure (Eq. 3). Under conditions of constant permeability and constant capillary pressure, Eqs. (3) and (7) can be combined to obtain the following relation-ship between infiltration length L, time t, and other proc-essing parameters: L = (1/e) [2ktWjS,p(l- e)/p.]"2 (8) Note that W; (work of immersion) is equal - yiv cosG. The Eq. (8) can be simplifted introducing that e"'V2kSsp(l- e) is the infiltrability of porous preform Q: L = Q [Wi/n]"2t1'2 (9) Again, it's important to note that Eq. (9) is valid under conditions of constant infiltrability and porosity of ceramic filler, constant work of immersion and, ftnally, constant viscosity of the melt, which is very difficult to obtain in practice. In spite of this considerable limitation, Eq. (9) can be successfully used in combination with Eq. (3) in order to designe the simple mathematical criterion for an early stage of pressureless infiltration of porous ceramic preform. Moreover, using this procedure, the parameters of pressureless infiltration can be selected to satisfied both proceesing requirements: spontaneous infiltration at ac-ceprtable infiltration rate. 3 Materials and experimental procedures Preparation of porous SiC preforms For the purpose of this study, three basic SiC mor-phologies - particles, platelets and vvhiskers in several size ranges (Table 1) vvere used for preforms preparation. Photomicrographs of used povvders are compared in Fig. 1. A diagram outlining the preform production process is shovvn in Fig. 2. Table 1: Characteristics of SiC phases used Particles Platelets Whiskers HSC 1200 SiC Platelets M-Grade SiCw Micro grits Millenium Advanced Superior Graphite Materials. Inc. Refractory Technologies, Inc. Chemistry: Stoichiometric SiC Stoichiometric SiC Stoichiometric SiC Crystalographic Primary phase Beta Primary phase Beta Primary phase Beta Structure: Diameter range 2-12 35-40 whisker length (Hm): 15-20 Thickness (um): / 3-5 1-2 Aspect ratio: / 8-10 10-12 Purity: 97-99 wt% SiC <1000 ppm of <1000 ppm of metallic impurities metallic impurities Particulate Content -100 5-10 5-10 (%)■■ Oxygen (%) by 1.0 0.68 1.1 Lečo Free Carbon(%): 1.0 0.01 0.53 Specific Gravity (g/cm3): 3.21 3.21 3.21 (7) Preform infiltration The experimental lay-up used in this vvork consisted of an aluminium alloy ingot, measuring about <|)50 x 30 mm, placed on the top of a porous ceramic preform. The filler material had a height that vvas great enough to pre-vent full infiltration under the process conditions (i.e. more-or-less infinite column of filler material). After processing, the amount of infiltration (distance from al-loy/filler interface) vvas measured, and the composite vvas sectioned and examined both macro- and micros-tructurally. The alloy/filler pairs vvere than placed into a controlled atmosphere furnace vvithin a refractory vessel (a 99.9% sintered alumina). The furnace vvas evacuated to -1 Pa at room temperature and back-filled vvith an ni-trogen-containing atmosphere until a positive flovv vvas obtained. Note that ali experiments vvere conducted under a slight positive pressure that vvas achieved by bub-bling the exit gas through a 25 mm column of oil. Fol-lovving the procedure developed in Lanxide, the furnace vvas ramped to temperature at a rate of 200°C/h, held at temperature for the specified time (e.g. at 800-1000°C for 10 to 24 h for full infiltration of the speciments) and allovved to cool to 675°C, at vvhich time the samples vvere removed from the furnace and cooled to room temperature. Various combinations of magnesium-containing aluminium alloys, silicon carbide porous preforms, nitro-gen-containing gases, and temperature/time conditions vvere employed to study the effect of various process variables on the infiltration kinetics. Because the infiltration of the porous preforms oc-curs in a nitrogenous atmosphere (at least about 10 volume percent nitrogen and the balance a non-oxidizing gas under the process conditions), aluminium nitride pre-cipitates may form vvithin the aluminium alloy matrix. Figure 1: Photomicrographs of used SiC morphologies: a) HSC 1200 Mierogrits Superior Graphite, b) SiC Platelets, Millenium Materials, Inc. and c) M-Grade SiC whiskers, Advanced Refractory Technologies, Inc. Slika 1: SEM fotografije SiC uporabljenih delcev: a) SiC prah -HSC 1200 Mierogrits Superior Graphite, b) SiC ploščice, Millenium Materials, Inc. in c) SiC vvhiskerji - M-Grade, Advanced Refractory Technologies, Inc. The per cent weight gain provides a measure of the amount of aluminium nitride that forms during processing. For comparison, the total conversion of pure aluminium to aluminium nitride produces a weight gain of 52%. Moreover, because this experimental arrangement produced a constant volume of composite in ali cases where full infiltration occured, the vveight gains of different experiments could be directly compared. 4 Results and discussion Preform preparation SiC preforms, containing vvhiskers, platelets or particles, vvere fabricated by vacuum casting in a variety of shapes and vvith a uniform microstructure. The character-istics of these preforms are listed in Table 2. Table 2: Characteristics of SiC- vvhiskers, platelets and particles preforms made by vacuum casting method CHARACTERISTIC PREFORM Particle's grade Platelefs grade Whisker's grade Average bulk density (g/cm3) 1-2.25 1-2.25 1-2.25 Preform diameter (cm) 3-10 3-10 3-10 Preform height (cm) 2-5 2-5 2-5 BET-Specific surface area (nr/g) 1.5-5.9 2.0-2.5 3.5-3.8 Porosity (vol%) 30-70 30-70 30-70 Infiltration experiments The critical process conditions for pressureless infiltration of porous SiC preforms vvith molten aluminium Ultrasonic homogenisation of raw materials (SiC phase + 5 wt% polyphenilene - corresponding to 4% cxcess carbon) in toluene i Evaporation of toluene vvith constant stining 4 _Addition of 1 wt% amorphous boron in the slip by stining_ _l_ Treatment of mould surfaces with an ammoniumalginate solution to enhance _mould release_ _i_ 1 Vacuum casting in plaster mould | _i_ L_Removal of preform from mould | _i_ f Diying of Preform at 75°C I _i_ | Pyrolysis of polyphenylene in Argon Flow ( 4h at 450"C ) | _i_ | Sintering of Porous Body at 1900 - 2000°C for 0.5 h in Argon ( 100 KPa) | _i_ Ojddation of Preform in Air at 900°C for 4 hours ____or Other Surface Treatment_ Figure 2: Preform fabrication process Slika 2: Proces pridobivanja poroznih predoblik Figure 3: Relationship betvveen magnesium content in an Al-lOSi-Mg alloy and iniltration distance (process conditions - 5 h dwell at 1175 K, nitrogen atmosphere vvith 4 vol% Ar) measured in SiC particle grade preform Slika 3: Odvisnost globine infiltracije od vsebnosti magnezija v Al-lOSi-Mg zlitini (eksperimentalni pogoji - 5 h pri 1175 K, atmosfera dušika s 4 vol% Ar) za porozne SiC predoblike pripravljene iz SiC prahu alloys vvere found to be: (i) the alloy composition, (ii) the atmosphere composition, (iii) the process temperature and time and (iv) the infiltrability of the preforms. The influenece of alloy composition (specifically the magnesium content) on infiltration distance is plotted in Fig. 3. The collected results are in agreement vvith data previously reported by Aghajanian et al.12 The nevv data also confirm the linear relationship betvveen magnesium content and amount of infiltration proposed by Aghajanian et al.12. The effect of nitrogen content of the atmosphere on the infiltration process vvas determined by conducting experiments in atmospheres ranging from 100% N2 to 100% Ar. It vvas found that no infiltration occured in 100% Ar, only partial infiltration occured in 10% N2+90% Ar and full infiltration occured vvhen the nitrogen content exceeded 20-30 vol%. As reported", at high percentages of N2, vvhere infiltration vvas rapid, little ni-tride formed, vvhereas in dilute atmospheres, vvhere infiltration vvas slovv, observable levels of A1N formed. In a similar fashion, the process temperature significantly af-fects the quantity of nitride that forms vvithin the aluminium alloy matrix. Figure 4 plots unit vveight gain versus process temperature for samples using alloy Al-10Si-3Mg, preforms made by SiC grit and process conditions of a 5 hour dvvell at temperature in 95% N2/4% Ar. Results also demonstrate that increased process temperature results in increased nitride formaton vvhichin-crease becames significant and nearly linear for temperatures higher than 1125 K. At a constant magnesium level and a fixed nitrogen content, several others process variables can affect the infiltration behaviuor. Fig. 5 plots the infiltration distance against temperature for otherwise constant process conditions. It is evident that infiltration increased in an approximatively linear manner vvith the temperature. Ad-ditionally, the data shovv that there is a threshold temperature required to initiate the pressureless infiltration for a given set of process parameters. Although limited, the data presented in Fig. 6 suggest that the threshold temperature changes vvith the process conditions (the preform infiltrability, the alloy composition and the nitrogen content in the processing atmosphere). One can also conclude that the process temperature affects the quantity of A1N that formed vvithin the aluminium alloy matrix. Fig. 7 plots the unit vveight gain against temperature for different SiC grade preforms. The results demonstrate that as the process temperature increases, the quantity of A1N that forms also increases. These results are in vvell agreement vvith reported data12 and confirm that the increase in A1N content is approxi-mately exponential over the temperature range investigated. 0.5 4 . O J-t-t--1-1-1-1-1-}-1 10 20 30 40 50 60 70 80 90 100 Nitrogen in atmosphere (%) Figure 4: Dependance of unit vveight gain on the content of nitrogen in N2/Ar atmosphere (Al-10Si-3Mg alloy, SiC particle grade preform, 5 h soakat 1075 K) Slika 4: Odvisnost povečanja teže vzorcev od vsebnosti dušika v N2/Ar atmosferi (zlitina Al-10Si-3Mg, predoblike pripravljene iz SiC prahu, 5 h pri 1075 K) eo j 70 i 10 Process temperature (K) Figure 5: Variation of infiltration distance vvith process temperature (Al-10Si-3Mg alloy, SiC particle grade preform, 5 h soak at temperature in a nitrogen atmosphere vvith 4 vol% Ar) Slika 5: Odvisnost globine infiltracije od temperature (zlitina Al-10Si-3Mg, predoblike pripravljene iz SiC prahu, 5 h pri delovni temperaturi v dušikovi atmosferi s 4 vol% Ar) Magnesium contant (wt%) 1400 -r 2 1000 -• 3 ; 800 -a E • Z 600 0 1 f 400 -200 ■• o -1-1-1-1-1-1-I 30 40 50 60 70 80 90 100 Nitrogan In itmoipKen (wt%) Figure 6a: Relationship between threshold temperature and infiltrability of porous SiC particle grade preforms (Al-10Si-3Mg alloy and a 5 h soak at temperature in nitrogen atmosphere with 4 vol% Ar) Slika 6a: Odvisnost temperature začetka infiltracije od infiltrabilnosti poroznih predoblik, pripravljenih iz SiC prahu (zlitina Al-10Si-3Mg, 5 h pri delovni temperaturi v dušikovi atmosferi s 4 vol% Ar) Figure 6b: Relationship between threshold temperature and magnesium content in Al-lOSi alloy (SiC particle grade preform and a 5 h soak at temperature in nitrogen atmosphere vvith 4 vol% Ar) Slika 6b: Odvisnost temperaturnega praga od vsebnosti magnezija v Al-1 OSi zlitini (predoblike pripravljene iz SiC prahu, 5 h pri delovni temperaturi v dušikovi atmosferi s 4 vol% Ar) Figure 6c: Relationship between threshold temperature and nitrogen in processing atmosphere (Al-10Si-3Mg alloy, SiC particle grade preforms and a 5 h soak time at 1125 K) Slika 6c: Odvisnost temperaturnega praga za porozne predoblike pripravljene iz SiC prahu, od vsebnosti dušika v delovni atmosferi (zlitina Al-10Si-3Mg, 5 h pri delovni temperaturi) Figure 7: Relationship between process temperature and A1N formation (unit vveight gain) in aluminium alloy matrix (obtained using alloy Al-10Si-3Mg, SiC particle grade preform and a 5 h soak at temperature in nitrogen atmosphere with 4 vol% Ar) Slika 7: Odvisnost temperature infiltracije od deleža nastalega A1N (izraženega kot povečanje teže analiziranih vzorcev) v Al zlitini (zlitina Al-10Si-3Mg, predoblike pripravljene iz SiC prahu, 5 h pri delovni temperaturi v dušikovi atmosferi s 4 vol% Ar) The effect of the infiltrability of porous preforms (see Eq. (9)) on the infiltration process vvas studied using preforms vvith different porosity and specific surface area. Note that the preform infiltrability, defined as a" 'a2kSsp(l-a), could be expressed as a function of specific surface area (Ss) and porosity (a) taking into ac-count Eq. (5): Q = const. e-'SsVp(l- e) (10) Fig. 8 plots the infiltration distance against preform infiltrability. The changes in the infiltrability of porous preforms vvere obtained by ranging their porosity and specific surface area. In order to meet these require-ments, the preforms vvere prepared using selected sintering conditions. The results demonstrate that ali experi-mental data fit vvell vvith the proposed process kinetics expressed by Eq. (9) for othervvise constant process conditions. Moreover, Eq. (9) seems to be valid for very different morphology of SiC particles. Hovvever, the Eq. (9) also, in some matter, presents a serious problem. There is a very complex correlation betvveen preform porosity and its real specific surface area. Usually, BET technique is used to determine Ss. It should be noted, hovvever, that a method based upon gas adsorption at the surface vvhose area is to be measured may not provide the right value to be inserted in Eq.(9). In fact, as reported14, the specific surface area relevant in the vvetting of particulates by aluminium could be much lovver than that given by the BET technique. 90 80 70 E S 60 0 | BO n 1 40 ■ I 30 20 10 0 80 70 10 0 E 50 E 1.5 2 2.6 Infiltrabilitv, Q 3 3.5 4 1.6 2 2.6 3 Infiltrabilitv. O 3.5 4 0 0.5 1 1 5 2 2.6 3 3.5 4 Infiltrabilitv, Q Figure 8: Dependence of infiltration distance on infiltrability of different porous preforras (Al-10Si-3Mg alloy, a 5 h soak at 1175 K a nitrogen atmosphere with 4 vol% Ar) for - a) SiC particle grade preform, b) SiC platelets grade preform, and c) SiC whisker grade preform Slika 8: Odvisnost globine infiltracije od infiltrabilnosti poroznih predoblik, pridobljenih iz: a) SiC prahu, b) SiC ploščic in c) SiC whiskerjev (zlitina Al-10Si-3Mg, 5 h pri 1175 K, dušikova atmosfera s 4 vol% Ar) 5 Concluding remarks A process for the production of porous SiC preforms consisting particles, platelets or whiskers is reported. It involves the vacuum casting of specialy prepared slip and sintering of green body to the porous specimen. Fol-lowing this procedure, vacuum čast preforms in a variety of sizes, vvith high dimensional and compositional repro-ducibility, and vvith uniform characteristics vvere fabri-cated. Porous preforms vvere sucessfully pressureless infil-trated using the Primex™ method originaly developed by Lanxide, Inc. The results presented in this article demonstrate that the combination of magnesium in the Al alloy, the nitro-geneous processing atmosphere (vvith at least 25 vol% N2) and several porous preform characteristics (specifi-cally the porosity and specific surface area) summarised in term preform infiltrability leads to the pressureless infiltration of molten aluminium alloy into ceramic filler. The collected data have confirmed that no infiltration occured vvithout the correct combination of above listed process variables. This means that the magnesium con-tent in Al alloy and the content of nitrogen in processing atmosphere should be combined vvith correctly designed porous preform characteristics. The results of present vvork demonstrate the infuence of preform porosity and its specific surface area on the infiltration length. The infiltration kinetics vvere shovvn to be strongly affected by preform infiltrability for other-vvise constant process conditions. In addition to kinetics, it vvas found that experimental data fit vvell the proposed expression for the infiltration length as a function of preform infiltrability, vvork of adhesion, viscosity of the melt and processing time. Moreover, the experimental results demonstated that the suggested equation is opera-tive for the different morphology of SiC particulate used in this vvork. Hovvever, the influence of preform surface composition, vvhich should affectsthe vvork of adhesion, and the viscosity of the melt is matter of the further ex-perimental efforts. 6 Acknowledgements The porous ceramic preforms have been infiltrated using the laboratory feasibilities of the partners on Brite Euram project BE96-3925 (proposal). The authors also thank to Slovene Ministry of Science and Technology, Impol from Slovenska Bistrica and Nova Kreditna Banka Maribor for their financial support. This research is the part of basic scientific project: "Preparation and Characterisation of Discontinuously Reinforced Al-SiC MMCs" fully financed by Slovene Ministry of Science and Technology. 7 References 1 D. J. Lloyd, Intern. Materials Reviews, 39, 1994, 1-23 2T. S. Srivatsan, I. A. Ibrahim, F. A. Mohamed, and E. J. Lavernia, J. Mater. Sci., 26, 1991, 5965-78 3 A. Mortnesen and T. Wong, Met. Trans. A, 21A, 1990, 2257 "F. Folgar, Ceram. Engng. Sci. Proc., 9, 1988, 579 5 J. T. Burke, M. K. Aghajanian and M. A. Rocazella, Proc. Int. SAMPE Symp., 34, 1989, 2440 fiJ. A. Comie, A. Mortensen and M. C. Flemings, Proc. of ICCM-VI (ed. F. L. Mathews, et al.), 1987, 2.297-2.319 7 S. Nagata, A. Kitahara, S. Akiyama, and H. Ueno, Trans. AFS. 93. 1985, 49-54 8 S. Caron and J. Masounave: in "Fabrication of particulates reinforced metal composites", (ed. J. Masounave and F. G. Hamel), Materials Park, OH, ASM International, 1990, 107-113 "K. G. Daviš and J. G. Magny„ Trans. AFS.. 89, 1981, 385-402 10 A. W. Urquhart. Mater. Sci. Eng., A144. 1991. 75 "M. K. Aghajanian, M. A. Rocazella, J. T. Burke and S. D. Keck, J. Mater. Sci., 26. 1991, 447-454 12 US Pat. 4 828 008 "J. W. McCoy, C. Jones and F. E. Wamer, Sampe Q„ 19, 1988, 2, 37 l4R. Sharma, Colloids Surf., 16, 1985. 87 15 L. R. White, J. Colloid Interface Sci., 90, 1982, 536 16 J. Bear: in Dinamics of Fluids in Porous Media, Dover Publication, New York, NY, 1988